3 ms·

Could you use a cake slicing algorithm? I.e. you get one party to choose a fraction, and the other party chooses whether it's theirs or not, until the number of

by veltas 3y ago

Could you use a cake slicing algorithm? I.e. you get one party to choose a fraction, and the other party chooses whether it's theirs or not, until the number of fractions is gone? Does this scale to n parties of different shares and allegiances splitting a 2d plane? Or maybe just generate the borders with a heat map after elections with PR?

Folcon 3y ago
Could do a variation, one party decides all the slices and then parties take turns picking in rounds, but the slicer always picks last.
- vlovich123 3y ago
  As the sibling to your comment noted that cake slicing only works like that if there’s a homogeneous preference for slices between parties but that’s not the case with districts. What I don’t get here though is what is the other party picking in this scenario. Is this saying in lue of voting just give both parties equal amount of districts to represent? Then what’s the point of cake slicing?
- bena 3y ago
  That simply does not work. The "slices" are done so that each slice have roughly the same amount of people. So if one party has sliced things so that all the slices are nicely cracked and packed, the "picking" just doesn't matter. The slicing is the picking. However, a variation of your idea is to have each party take turns drawing districts. But that doesn't really work, because you'd quickly learn that the winning strategy is to draw districts to maximize the other party's voters. Because you want to force the other party to draw districts favorable to you.
tetrep 3y ago
You can't apply cake slicing logic because in that scenario everyone likes the (implied homogeneous) cake equally. With gerrymandering, how we slice the cake determines how appealing it is to each party involved, so the gerrymandering slicer could slice a district that would never appeal to the other party. The other party would of course reject it, so you'll at best be in a deadlock. Whatever solution you might implement to avoid a deadlock can't detect which party is being "unreasonable" in causing the deadlock, so you would need to have another different way of drawing districts, and whatever method that is is likely to be affected by gerrymanderers too. Even something basic like using a uniform 2D grid would likely bias one party or the other, giving an incentive for that party to desire a deadlock for any reason, as they would then win by default.
penteract 3y ago
If carefully done, something like this (where parties play a game to decide the borders) might avoid partisan gerrymandering, but there's also reason to be concerned about bipartisan gerrymandering where elected representatives from 2 (or more) parties prefer to have safe seats, so they agree to group voters with similar ones in a way that means it takes a very large swing for anyone to lose a seat and change the result.
MengerSponge 3y ago
Yes, and it works really well for two parties if both parties are equally informed and competent. https://arxiv.org/abs/1710.08781 https://arxiv.org/abs/1710.08781 For n>2 the simple I-cut-you-choose stops working.
brianolson 3y ago
For many years California had a bipartisan gerrymander that made the 60:40 D:R state have about 60/40 in its state legislature and US House districts and every seat was a safe seat. That's a certain kind of democracy death. :-/ I'd much rather have any algorithm that gives no power to any parties, and solves a better answer to the question "what is a district for?" (my short short: Local representation for local issues) As other comments have also noted, lots of people don't want _geographic_ representation but some kind of other issue/identity Proportional Representation. Yup. Also true. My dream State Legislature has a PR Senate at-large across the whole state in batches of 20 or so, and a small-local-district House.

by veltas 3y ago

Folcon 3y ago

Could do a variation, one party decides all the slices and then parties take turns picking in rounds, but the slicer always picks last.

vlovich123 3y ago
As the sibling to your comment noted that cake slicing only works like that if there’s a homogeneous preference for slices between parties but that’s not the case with districts. What I don’t get here though is what is the other party picking in this scenario. Is this saying in lue of voting just give both parties equal amount of districts to represent? Then what’s the point of cake slicing?
bena 3y ago
That simply does not work. The "slices" are done so that each slice have roughly the same amount of people. So if one party has sliced things so that all the slices are nicely cracked and packed, the "picking" just doesn't matter. The slicing is the picking. However, a variation of your idea is to have each party take turns drawing districts. But that doesn't really work, because you'd quickly learn that the winning strategy is to draw districts to maximize the other party's voters. Because you want to force the other party to draw districts favorable to you.

tetrep 3y ago

You can't apply cake slicing logic because in that scenario everyone likes the (implied homogeneous) cake equally. With gerrymandering, how we slice the cake determines how appealing it is to each party involved, so the gerrymandering slicer could slice a district that would never appeal to the other party. The other party would of course reject it, so you'll at best be in a deadlock. Whatever solution you might implement to avoid a deadlock can't detect which party is being "unreasonable" in causing the deadlock, so you would need to have another different way of drawing districts, and whatever method that is is likely to be affected by gerrymanderers too. Even something basic like using a uniform 2D grid would likely bias one party or the other, giving an incentive for that party to desire a deadlock for any reason, as they would then win by default.

penteract 3y ago

If carefully done, something like this (where parties play a game to decide the borders) might avoid partisan gerrymandering, but there's also reason to be concerned about bipartisan gerrymandering where elected representatives from 2 (or more) parties prefer to have safe seats, so they agree to group voters with similar ones in a way that means it takes a very large swing for anyone to lose a seat and change the result.

MengerSponge 3y ago

Yes, and it works really well for two parties if both parties are equally informed and competent. https://arxiv.org/abs/1710.08781 https://arxiv.org/abs/1710.08781 For n>2 the simple I-cut-you-choose stops working.

brianolson 3y ago

For many years California had a bipartisan gerrymander that made the 60:40 D:R state have about 60/40 in its state legislature and US House districts and every seat was a safe seat. That's a certain kind of democracy death. :-/ I'd much rather have any algorithm that gives no power to any parties, and solves a better answer to the question "what is a district for?" (my short short: Local representation for local issues) As other comments have also noted, lots of people don't want _geographic_ representation but some kind of other issue/identity Proportional Representation. Yup. Also true. My dream State Legislature has a PR Senate at-large across the whole state in batches of 20 or so, and a small-local-district House.